For our 2x2 experiment (4 groups), we found that those in condition A did recognize their manipulation accurately, but that there is a spillover effect and the two groups in condition B also report different levels of A.
- 2x2 scenario experiment on e-mails
- condition 1 content positive vs negative
- condition 2 tone polite vs impolite
What we see is that the two groups in positive content rate their email as more positive in content than those in negative content - so manipulation successful. But, those in polite tone also rate their e-mail als more positive in content than those in the impolite tone condition. But the F-values are quite far apart.
How do I show that despite this spillover, the F-value of the manipulation of content is so much greater (signif) than the F-value for the tone groups?
I want to show that these two F-values differ significantly - right? Using SPSS, could anyone recommend the steps to take here. I've been recommended a Wald test but am struggling.